Specific Aims: To examine the P300 and N4OO Event-Related Potentials (ERPs) in a well-characterized cohort of older schizophrenia patients, with comparably aged samples of patients with late-onset schizophrenia (LOS) and early- onset schizophrenia (EOS). P300 and N400 abnormalities will each be related to specific clinical, neuropsychological. and neuroanatomical (ME volumetry) measures. if, as we hypothesize, different patterns of ERP abnormalities are found in EOS and LOS, this win advance our understanding of "the schizophrenias",(i.e. a likely heterogenous group of disorders). Our preliminary data suggest relative sparing of the P300 amplitude in LOS, while the N400 latency is somewhat more delayed in LOS than in E05. Main Hypotheses: I) The P300 amplitude will be reduced in EOS patients, compared to age-matched normal controls and LOS patients. Smaller P300 amplitudes will be related to impairment in attention and learning, and to smaller hippocampal and thalamic volumes. 2) N40O amplitude reductions will occur predominantly in EOS patients. The N400 amplitude will be reduced in patients characterized by more severe negative symptoms (especially poverty of thought and alogia) and abnormal language abilities. The secondary hypotheses include: Schizophrenia patients with abnormal cognitive ERPs (P300 or N400) will have poorer global cognitive function than patients with normal ER Pg. Methods: Thirty subjects from each group (EOS, LOS, controls) will be selected from the UC San Diego Geropsychiatry Clinical Research Center with comparable age, education, ethnicity and gender. All subjects will have had undergone recent clinical assessment (within 7 days), MRI scanning and extensive neuropsychological testing. They will be tested on ERP paradigms designed to elicit the P300 (i.e. an auditory target detection task) and N400 (i.e. category and antonymic decision tasks) potentials. Twenty-one channels of EEG data will be collected and undergo ERP analysis, defining the P300 and N400 amplitudes and latencies by standard protocol. To test our main hypotheses, intergroup differences of the P300 and N40O amplitudes will be tested by using split-plot ANOVAs (group x electrode site). These ERP amplitudes will each be modeled as a function of specific clinical, neuropsychological and MRI volumetric measures using correlational and multiple regression analyses. Long-term Objectives: If this proposal were funded, it would allow the direct comparison of the P300 and N400 in an older cohort of schizophrenia patients. which is of great theoretical interest. This study would help define the clinical significance of specific ERP abnormalities with respect to symptomatology and type of cognitive deficit. It will further our understanding of late-onset schizophrenia. Future longitudinal studies would allow the etectrophysiologic quantification of cognitive decline and the testing of to what decree cognitive ERP abnormalities predict floor prognosis.